The algorithm completed matching processes in 64 seconds, far faster than the top competitor’s time of 43 minutes and the test average of 1 hour and 42 minutes. ROC’s technology showed a 30.3 percent accuracy improvement over 12 months, with the False Negative Identification Rate (FNIR) decreasing from 0.198 with ROC +0009 to 0.138 with ROC +0014.
Dr. Brendan Klare, Chief Scientist and Co-Founder of ROC, noted the exceptional speed of the system: “Very few providers can match a 30 million fingerprint database in 1 minute,” highlighting the advancement in relation to recent developments in forensic fingerprint technology.
The NIST ELFT benchmark tests searches using images and feature sets marked by human latent fingerprint examiners. ROC’s evaluation used the same algorithms available in their SDK, which the company provides for free testing. This capability supports new methods for accelerating forensic investigations.
Based in Denver, CO, with facilities in Morgantown, WV, and Grand Rapids, MI, ROC is the only American-made multimodal biometrics and computer vision provider. The employee-owned company develops technology for on-device and on-the-spot matching during crime scene investigations, advancing the field of mobile identification solutions for law enforcement.
—
February 20, 2025 – by the ID Tech Editorial Team
